Immunology of Inflammatory Bowel Disease and TNF Inhibition

Grand Rounds

Lloyd Mayer, MD

Professor of Medicine
Mount Sinai School of Medicine

What I would like to do over the next 45-50 minutes is basically take you through a tour of the current knowledge base in the immunopathogenesis of inflammatory bowel disease and take the leap from our newer understanding of these diseases into what has become an incredibly fast paced field in novel therapeutic approaches.

I will start with the first slide that comes from a very recent review from Joe Kirsner in Gastroenterology three years ago, which talked about at that point in time the current etiologic theories of Crohn's disease. What you see here is that there are a huge number of different possibilities in terms of what might be triggering disease onset. What has really been reinforced over the last three years, is that this broad list of potential etiologic mechanisms have been boiled down to three interacting phenomena, which are seen in this triangle. In a genetically predisposed host exposure to some environmental trigger results in an abnormal, immune response which leads to nonspecific inflammation. The body's attempt to control this inflammation causes some tissue injury. Then, most importantly restitution and repair processes take hold. Each one of these processes in inflammatory bowel disease appear to have gone awry. So hopefully I will be able to convince you of that and tell you where we can now, using this knowledge, intercede with novel approaches to therapy.

I mentioned that there are environmental triggers and we have actually known this for the past 50 years. We have case histories where patients come into the office after having traveled to Mexico or to the Middle East or even sometimes to the southwest and have developed some form of turista. A husband and wife traveling together develop turista with a diarrheal syndrome. The husband and wife are given some Pepto-Bismol to treat the infection. The husband gets better but the wife continues to have diarrhea, then develops bloody diarrhea and is seen by her gastroenterologist and diagnosed as having ulcerative colitis.

So there are clear infectious triggers, and I will really underscore this later on, that can start the ball rolling in terms of development of these disease processes. There are a huge group of patients who, following use of NSAIDs for some joint symptom, will develop diarrheal syndrome, eventually bloody diarrhea, or abdominal cramping, and will be diagnosed as having ulcerative colitis. Clearly post-antibiotic induced diarrhea, not necessarily C. difficile induced diarrhea, can trigger the initiation of these diseases.

Now, what you see at the bottom of this slide: psychological distress and diet are aggravating conditions for development of inflammatory bowel disease. People have not really bought into the fact that they actually induce the disease to begin with, but clearly they can aggravate an already existing condition. The other interesting phenomenon is actually a dichotomy between the two diseases of inflammatory bowel disease, ulcerative colitis and Crohn's disease, as the disease association with smoking. Stephen Hanauer and the University of Chicago, probably about eight to ten years ago, had described the typical presentation at least in the Chicago area, of patients who are long-term smokers who had been advised to stop smoking and within about three to twelve months following cessation of smoking would develop a diarrheal syndrome and be diagnosed with ulcerative colitis; and if these patients went back to smoking or they started taking Nicorette Gum or Nicotine Patches that their bowel disease would come under some form of better control. So this is clear association that has been confirmed by a number of centers and people have tried desperately to figure out what within cigarette smoke is actually inducing this sort of calming effect on the bowel. As I will allude to later, the recent data suggests that in fact this relates to the barrier function of the intestine being maintained by either nicotine or some nicotine derivative. That is quite intriguing in terms of our understanding of these diseases, and I will point out why that is a little bit later. Interestingly, in Crohn's disease, the opposite is the case; smoking actually can aggravate Crohn's disease, and we do not really understand why that is.

Now this brings us to sort of a coordinated approach. I mentioned before that there maybe some infection. There has been a lot of work in this area for years and years; as a matter of fact, that is why I avoided IBD research for many years when I first went to the lab at Rockefeller. I did not want to touch IBD with a ten-foot pole because the major focus had been that these diseases were mediated by a specific infection, not unlike the studies that have been done in rheumatoid arthritis. There is a large literature on the role of mycobacteria. This hypothesis was just resurrected a few months back in the L.A. Times saying that you should not drink cow's milk because you are going to get a mycobacterial infection that is going to cause Crohn's disease. There are other groups that have thought that there is some Helicobacter species, which triggers the onset of inflammatory bowel disease. There are groups in England that seem to be facile in finding these new agents. They have either felt that there are some specialized types of bacteria or L forms or alternatively measle virus. Wakefield has been standing on a soapbox claiming that this is the causative agent of Crohn's disease. I think the data to date strongly suggests that none of these specific infections are in fact specific for inflammatory bowel disease but like the example that I gave you with turista or post-antibiotic induced C-difficile colitis, or just post-antibiotic induced diarrhea, that there are some exogenous agents that can trigger the disease but do not necessary mean that they are specific for the disease.

Now, what happens is you are exposed to some initiating event and then either because of a genetic predisposition in terms of a defect in the barrier function, (I alluded to that in terms of the role that nicotine may play in helping to maintain barrier function or mucin production) or in the mucosal immune system causing the loss of tolerance to things that we put into our intestines actually causes the disease. This approach has led to what people have felt to be a coordinative hypothesis of the disease. I can read this off, and sort of emphasize the main points - That chronic intestinal inflammation and related systemic manifestations are due to an overly aggressive immune response to resident luminal bacterial constituents, and I will allude to that a little bit later as well. They appear to be at least for Crohn's disease mediated by a specialized subpopulation of T helper cell, Th1 cell, as well as macrophages. The genetic factors as well as the factors involving barrier and immune responses are clearly predisposing and that there has to be some initiating trigger. This is the hypothesis that people have put forward.

Why do we think that the immune system is involved? Here is where we got involved and this is where I think the biggest bang for the buck has occurred. This is a histologic picture from a patient with Crohn's disease and this is a granuloma, the hallmark of Crohn's disease, although we generally see this hallmark as a non-caseating granuloma in only about 10 to 20 percent of patient biopsies. So you can see here that to a gastroenterologist, this granuloma is Crohn's disease but to immunologist or immunopathologist this is the hallmark of a delayed type hypersensitivity reaction or a Th1 mediated immune response.

In contrast, look at the picture in ulcerative colitis and the picture histologically is quite different. You have this very aggressive inflammatory response, you have crypt abscesses here; these are the crypts over here. These crypt abscesses filled with neutrophils, a lot of mass cells present within the inflammatory response, and this is what an immunologist would call an Arthus reaction or immune complex mediated tissue injury. Very different from the very localized and less destructive type of response that you see in Crohn's disease.

Well, in order to understand what is going wrong in inflammatory bowel disease we have to understand what is going right in the normal mucosal immune system. I think the biggest hurdle in that field has been trying to sort of pull apart or tease apart immune responses that occur in the gut as opposed to those immune responses that occur systemically. I think the biggest difference has been our understanding that the type of immune response that is generated with an antigen that is delivered via the GI tract, is in fact focused totally differently. There are a number of unique phenomenons that are associated with mucosal immunity, oral tolerance, controlled inflammation, systemic suppression of immunoglobulin isotypes, with the exception of secretory IgA. All of these phenomena have one linked event, on the bottom here as suppression or down regulation of immune responses. This is the antithesis if what you see systemically if you inject an antigen under the skin or in the muscle. This focuses the immune response and tries to generate an aggressive immune response to eliminate the pathogen. But in the gut, which if you think about it, is essentially a sewer, you have 1012 to 1014 bacteria per gram of tissue. That is a huge antigenic load and that is just normal -- the good guys. There are good bacteria that are sitting there and you should not try and generate immune responses against those otherwise you would be chronically inflamed. So that is part of the problem, the immune system in the gut has to recognize who the good guys are, who not to respond to, and who the potential pathogens are that would allow you to break this tolerance type of phenomenon and induce a protective immune response. Well, the hallmark of this suppressed or regulated immune response is what I mentioned before as controlled or in terms physiologic inflammation.

This is a biopsy from a normal bowel. Here is the outside, the lumen out here, and what you see down here are crypts, the epithelial cells lining these crypts, and just one cell layer away from this sea of bacteria in the lumen is a mass of lymphocytes and they are not just lymphocytes, these are activated cells that actually constitutively produce fair amounts of cytokine and they are there constantly presently just churning out cytokines, producing antibodies in the case of the IgA plasma cells. Now, this picture in the gut will be called normal by a pathologist but you take this picture and you transplant it into a liver or a pancreas or even a joint and they are going to call this a chronic inflammatory disease. So it has to be underscored that we are dealing with a different level of mucosal immunoregulation when dealing with the GI tract.

So how can we take this and put it into the context of a better hypothesis of the development of inflammatory bowel disease? Well, here as you see all the way over on your left, the normal gut in pink is what we are calling controlled physiologic inflammation and anyone of us who gets exposed to an environmental trigger such as a viral gastroenteritis or Shigella, Salmonella, etc, will develop some degree of inflammation in response to those pathogens. However, in the normal setting there is a very rapid response in terms of trying to restore the integrity of the bowel so that factors come in to play that very quickly down regulate or dampen that immune response that allow you to bring things back down to this controlled physiologic inflammatory state. The hypothesis that has developed in inflammatory bowel disease is that this off switch is stuck in the on position. So it is sort of like the Energizer Bunny, it just keeps going and going and going, and you have this chronic inflammatory process, which in fact is IBD. Well, we know a lot of things now especially over the last 8-10 years in terms of who the culprits are in this switch defect. It turns out that it is the CD4 positive helper T cell, I alluded to this before, and in fact depending on the subpopulation of CD4 positive T cell that is activated you will get a histopathologic picture that is more consistent with Crohn's disease, a macrophage activating gamma-interferon, TNF alpha type of response that one sees, which causes localized inflammation and tissue destruction. In contrast, in patients with ulcerative colitis there tends to be production of cytokines which would promote more of the antibody mediated types of immune responses, potentially inducing autoantibodies, potentially inducing immune complex formation, and nonspecific tissue inflammation, neutrophil infiltration and transmigration, and this would lead to this Arthus type reaction. So we are dealing with two different types of immune responses yet triggered potentially by very similar pathogens.

Well, as I just alluded to culprits in this whole process are cytokines produced by these CD4 positive T cells and this is a slide, which is very familiar to all of you. This is the classical pathway of T helper cell activation and cross regulation. The T helper 1 cells being more pro-inflammatory, those that attribute to patients who develop Crohn's disease with IL-2 and TNF alpha production promoting this localized inflammatory response. This has been brought in to question more recently in ulcerative colitis a more Th2 dominant type of response is seen in ulcerative colitis, and (it is thought that these may be in fact cross regulatory), although data emerging now suggest that this paradigm, at least in humans, is falling apart quite rapidly.

Well, the sort of nail in the coffin that pins down cytokines as the major culprit in this colonic inflammatory disease, was the finding of spontaneous animal models of inflammatory bowel disease. This was a huge boon to the field of IBD research and it was totally fortuitous. A variety of molecular immunologist were very interested in the role that cytokines played in either T cell development of systemic immune responses. What they did was to generate a series of mice with targeted gene deletions for a variety of different cytokines. What they were looking for was essentially a boy in the bubble type scenario, thinking that they would induce cytokine defects and then immunodeficiencies, but what they found very interestingly was that these mice were born normally, survived normally, and grew normally until about 3 months of age. These mice then started developing a wasting syndrome, developed diarrhea, rectal prolapse, and they said okay we finally have an the infection and obviously it was going to come through the GI tract first anyway, here we go, we have animals with immunodeficiency. They cultured these animals upwards, downwards, and sideways and could not find any specific infection that was present in the gut but what they did find when they sectioned the bowel was marked chronic inflammation which was very reminiscence of inflammatory bowel disease. It was not exactly Crohn's disease, it was not exactly ulcerative colitis, it was more right-sided type disease, cecal disease and early transverse colon. They were sort of struck by the fact that this was a model for inflammatory bowel disease.

The fortunate thing for us was the fact that not all targeted gene deletions resulted in inflammatory bowel disease that it turned out to be very selective in terms which mice developed IBD and which did not. The overriding theme was that in fact if you developed a mutant mouse model that had a defect in the way the T cells communicated, either with themselves or with tissue macrophages or if you perturbed the way the T cells regulated the immune response that the default pathway was mucosal inflammation, not systemic immunodeficiency. This sort of made sense to a number of people who were involved in the field, because as I showed you before you are walking a tightrope in the mucosal immune system. You are antigenically challenged as opposed to what occurs in the systemic immune system, and so if you tip that balance in some way the default pathway is going to be mucosal inflammation.

Here you see the number of models, and this list is huge now. There are probably about 18 different models of inflammatory bowel disease that relates in some way to an imbalance in T cell regulation. Here you just see the ones that were originally described, the TCR alpha knockout, the IL-10 knockout, and the IL-2 knockout, all of these told us that antibodies had nothing to do with this disease. There is a skid mouse model where this transfer of Th1 cells in to a skid mouse and these mice develop full blown inflammatory bowel disease without any B cells in the system. You could breed all of these animals back into a B cell knockout phenotype and the disease marches right along.

So it tells us that antibodies are really not critical to the development of this disease. It also tells us this is more colonic inflammatory disease as opposed to an autoimmune disease because there is no real evidence for auto reactivity in these diseases. Where the reactivity lies is quite interesting, it is the enteric flora (the normal gut flora). What a variety of investigators have done -- and this was best done by Ed Ballish in Wisconsin and Balfour Sartor down in UNC -- was that they derived a number of these mouse models in germ free environments and low and behold despite the fact that they had that same genetic defect these animals did not develop the disease. They were fed the same mouse chow, a radiated mouse chow, as their littermates that were grown in normal conditions, in normal germ containing conditions.

The other important piece of this, as I alluded to before, is the role that the barrier plays. Normally you look at the gut and it is covered with a very thick mucous coat and the epithelial cells in the gut are linked together by tight junctions, very strong physical barrier to prevent access of antigens. If you disrupt this barrier by a number of different genetic manipulations, either inserting a (this is Jeff Gordon's work) dominant negative N-cadherin gene, which essentially loosens the tight junctions or deletion of one of the mucin type products called trefoil factors in the gut, barrier is disrupted. These animals are intact immunologically yet they develop inflammatory bowel disease. So either barrier dysfunction with colonic stimulation of the immune system will trigger the disease or alternatively a defect in the way in which the immune system handles the normal flora will result in intestinal inflammation.

Now the other key feature to these animal models has been understanding of how genetics plays a major role and sort of underscores the fact that IBD, like asthma, like RA, like diabetes, are in fact multigenic disorders. What has been done is they have taken the mice with defects in the IL-2 gene, IL-10 gene, whatever and crossed these mice into several different strains. What they find is that in some strains with the same genetic defect, they will develop severe colitis and in other strains the same genetic deletion will cause no colitis whatsoever. This tells you that there is a strong role for a lot of background genes or modifying genes that are present in different strains and it makes our life incredibly difficult in terms of understanding the genetics in humin IBD, where obviously we are outbred.

This leads us to bring back our model, because the mouse models really encompass all three parts of that initial triangle or pyramid that I showed you before. There is an immunologic defect in terms of the cytokine gene defect that has been introduced in these animals. Clearly that is a genetic predisposition to these diseases and there has to be some environmental trigger, which is normal flora to start the process going. So all of these being somewhat abnormal in these mouse models results in inflammatory bowel disease.

But why do we not all develop inflammatory bowel disease? It gets back to what I talked about before, this normal resistant state of controlled inflammation. We get exposed to dietary antigens, bacteria flora, variety of bacterial products, chemotactic peptides, and etc, and the physical barriers that we have, the mucous barrier, the tight junctions play a role in excluding a lot of these antigens from gaining access to the underlying mucosa immune system. Even when we do have antigens traveling across the epithilium we actually elicit a response which turns out to be quite regulatory or dampened. This has been shown to be mediated through two novel populations of CD4 positive helper T cells, (which may not be helper actually - it is probably a poor use of the term) or what we will call regulatory cells.

One which has been championed by Howard Wiener called the Th3 cell, which produces a potent immunosuppressant cytokine TGF beta upon activation or the other population which has been described out of the SCID mouse model and actually out of a human, graft versus host disease model in the gut so called TR1 or T regulatory 1 population of T cells which produce predominantly IL-10 and will shutdown T cell activating responses. What normally happens is that the normal constituents in the bowel will selectively trigger these regulatory T cell populations and not trigger the more inflammatory or pro-inflammatory types of responses that would result in inflammatory bowel disease.

So, how is this occurring? I will take a little sidetrack here and bring in some of the work that has been going on in our laboratory because we think this is actually quite exciting in terms of relating to what potentially may be going on in normal inflammatory bowel disease. Well, I will step back a second and bring you back and focus you in on the intestinal epithelial cell. For years and years people thought this was just a passive barrier. It was going to exclude macromolecules, it was not going to let bacteria get into the underlying mucosa associated lymphoid tissue and that was it. It was not a key player in anything. A number of laboratories including our own thought that this may not be the case because intestinal epithelial cells in fact express a lot of molecules which were associated with immune regulation, HLA class II molecules, class I molecules, and a number of nonclassical class I molecules and there are studies from a number of laboratories that showed that intestinal epithelial cells could take up soluble protein antigens which normal if taken up systemically would elicit an immune response. So we asked the question whether intestinal epithelial cells could stimulate an immune response despite the fact that these cells constitutively express MAC class II molecules, which would generally elicit a T helper 1 or 2 type of response and what we found was that antigen sample by intestinal epithelial cells activating peripheral T cells would induce the expansion of what turned out to be antigen nonspecific suppressor regulatory cells. Instead of being CD4 positive, these cells were CD8 positive.

This was a little bit of a puzzle to us and to really cut to the chase here we tried to figure out what was mediating the interaction between T cells and the epithelial cells that was resulting in the production of these CD8 positive suppressor cells. We made a series of monoclonal antibodies against normal epithelial cells and asked these antibodies to block the induction of these CDA positive regulatory cells and we came up with two monoclonal antibodies, B9 and L12. Again, to make a long story short both of these antibodies recognized a 180 kilodalton glycoprotein which we call gP180. Just to take you through what we know about gP180 to this point, as I mentioned it is a 180 kilodalton glycoprotein expressed by normal epithelial cells, 50% of this molecule is N-linked sugar, so it is very heavily and glycosylated. Both antibodies that you see on the bottom (B9colL12) recognize the carbohydrate side chain, which turned out to be critical regulatory parts of these molecules. There is an apically sorted GPI anchored form of this molecule, which is expressed in the lumen and there is a basolaterally sorted transmembrane form, which we think interacts with either cells within the epithelial space, intraepithelial lymphocytes, or in the lamina propria.

We have cloned this molecule and it turns out to be an alternatively glycosylated form of the old tumor antigen, carcinoembryonic antigen or CEA. Now functionally this molecule turns out to be incredibly interesting because this molecule has the capacity to bind to CD8 on CD8 positive T cells and is intimately associated with the activation of these CD8 positive T cells triggering a signaling pathway involving a variety of Src kinases including p56lck. It also forms a complex with class 1b molecules. So these are molecules that generally are not ubiquitously expressed. They are also actually expressed in the placenta, the other area where you would want suppression to occur. CD1 is a class Il molecule which is incapable by itself to bind to CD8 but it forms a complex with gP180. It actually makes the whole complex look more class 1 like and allows certain regulatory T cells to get activated. As I said, it does induce the activation of regulatory T cells and interestingly this molecule does not induce the production of suppressive cytokines but rather stimulations the production of gamma-interferon. We are not exactly sure why this is at this point.

Okay, so here is a molecule, which we think is intimately associated with the activation of regulatory cells and we think this is what is going on in the GI tract. Here is a stain from a normal colon and what you see here is the epithelial cell is diffusely stain, that this apically sorted form here a more railroad track type form, and then this basolaterally sorted form. Here is the lamina propria down here and there are processes extending from the epithelium into the lamina propria that allow lymphocytes to sit around the epithelium. If there is antigen that is coming in from the lumen through the epithelial cell you can envision that antigen would be processed, could potentially be complexed to class 1b molecules such as CD1d and the complex of CD1d and gP180 will be expressed in the basolateral surface or in the epithelial space and this would trigger these regulatory C cells allowing for what we see as controlled physiologic inflammation.

Well if that is the case, what is the story with inflammatory bowel disease where we do not see this regulation? Well, here is a series of studies that were done by Dave Eisenhardt in our lab. Here is what I showed you before, normal intestinal epithelial cells either from an individual with cancer or colonic inertia, or conversely the isolated epithelial cells or cells from patients with inflammatory, non-IBD control such as diverticulitis, ischemic colitis. If you culture peripheral T cells or lamina propria T cells with these epithelial cells and trigger them with some antigen, you will get this activation to CD8 +T cells as I showed you a couple of slides ago. In contrast, if you take epithelial cells from a patient with inflammatory bowel disease either ulcerative colitis or Crohn's disease and you take it from the area that is actively inflamed but more importantly if you take this from an area where there is no histologic evidence of inflammation, that occurs commonly in Crohn's disease which occurs in segments, what you see is a failure to activate these CD8 cells and rather see the induction of CD4 cells which produce a large number of cytokines including gamma-interferon, TNF, IL-2, etc. We asked the questions to why this is the case and it actually took us about four years to figure this out until we got the antibodies to gP180.

If you recall a couple of slides ago, I showed you the immunohistochemical (mab BA) stain of a normal colon. This is a biopsy from a patient with ulcerative colitis. You see quite strikingly here that there is no staining at the basolateral portion of the epithelium -- this is an uninvolved segment from a patient with ulcerative colitis. There is a defect in the expression of the basolaterally sorted form. The apical form, the GPI anchored form is normal. The basolateral form is the form that is going to interact with the lymphocytes in the lamina propria or in the epithelial space. If that is absent potentially you would not be able to generate a regulatory CD8 positive T cell response.

Here is the picture in patients with Crohn's disease. You actually do not see any expression of gP180. This is again a non-inflamed area of a patient with Crohn's disease. It is a different picture than what you see from ulcerative colitis. If you pull this all together and you look at freshly isolated cells that you can physically remove from intestinal resections, we can run them through a flow cytometer, stain them with the anti-gP180 antibodies, we can get a picture for how much expression there is or how much non-expression there is. You see here, this is normal IECs all the way on the left in the white squares and there is wide diversity of expression of this molecule on normal epithelial cells. In patients with Crohn's disease, which you see here on the right-hand side, whether the epithelial cells are isolated from an area of involvement or non-involvement, there is no or very little expression of this molecule. In patients with ulcerative colitis where you take this specimen from an area where there is no inflammation such as the right colon in patients who have left-sided disease or the ilium in patients who have pan colonic disease, there is also limited or low expression, very low expression of gP180. If you take it from an area where there is active inflammation in a patient with ulcerative colitis, you see an increase in the expression but it is only of the apically sorted GPI anchored form. The basilateral form is defective. So we think from this that there are two different defects in the expression of this regulatory molecule. In patients with ulcerative colitis there is a defect in the sorting or the essentially putting on a transmembrane tail on to this molecule; and in Crohn's disease we thing there is an intestine specific defect in the expression of this molecule. Each on of these defects could potentially result in the failure to activate CD8 positive T cells.

Well, the question is, is this genetically regulated? Fortunately, Lisa Toy in the lab was able to look at a series of monozygotic twins who were discordant for Crohn's disease or ulcerative colitis. These are two twins that were discordant for ulcerative colitis, monozygotic twins, and she stained the epithelial cells isolated from the rectum of both of these patients for gp180, and what you see in the green circle there, with the little blue arrows, is that there is a defect in the expression of gp180 in both individuals. This tells us two things; one - and this was true for the three monozygotic twin pairs that she studied - the defect in expression of gp180was concordant even though the disease was discordant. The defect in gp180 may be genetically regulated but more importantly the defect in gp180 alone is not sufficient to cause the disease. There has to be some other trigger that allows this disease to become expressed.

The other interesting feature of this also done by Lisa was looking in both patients with ulcerative colitis and Crohn's disease at other sites where gp180 is expressed. It is expressed in the placenta. Another site where you would want to suppress immune response, but it is hard to get placentas from males and it is hard to get a lot of placentas from pregnant IBD patients, so the next best thing was to look in the buccal epithelium, which we fortuitously found expressed gp180 as well. What you see here is a flow cytometric analysis of normal buccal epithelium, BE, and this very nice peak shift here staining with gp180 in the normal. This is compared to the intestinal epithelial cells in that same individual. There is more gp180 expressed in the intestional epithelium than in the buccal epithelium. In Crohn's disease despite the fact that there is no gp180 expressed in the gut it is expressed normally in the buccal epithelium suggesting that there is an intestine specific defect in Crohn's disease. In ulcerative colitis interestingly there is actually a reduction in gp180 expression in the buccal epithelium. So it suggests there may be epithelial cell specific defect within these two diseases that is present only in the gut.

How can we actually test this and see whether these defects that we have defined immunohistochemically are relevant in vivo? Well, we have to sort of put this in the context of controlled inflammation or tolerance induction so we decided with Lisa Chan in the lab to do a series of experiments looking for the ability of patients with inflammatory bowel disease to tolerize to a neoantigen that was given orally. The hypothesis that drove this experiment was that if these regulatory suppressor T cells are responsible for oral tolerance, and there is data to support that, as we have shown that suppressor mechanisms are defective inflammatory bowel disease then we should not be able to induce oral tolerance in these patients. We should see active immunity to a foreign protein that is administered orally. Here is the concept of oral tolerance, which actually dates back to H.G. Wells. The upper panel depicts systemic immunization in the mouse immunized with egg white albumin on day 1. On day 21, this mouse will develop both antibodies and T cell reactivity to egg white albumin or ovalbumin. In contrast if you first feed this mouse ovalbumin and then you immunize this mouse on day 21, day 24, day 35, whatever, this mouse is resistant to immunization, that is the definition of tolerance. This animal does not make T or B cell responses to this orally administered antigen. We know this is an active process and there are actually several mechanisms to induce tolerance but we know that we can transfer T cells from the tolerized mouse to a naïve mouse and that mouse will have tolerance transferred for that specific antigen.

Okay, so here is what we did. Actually these studies have been expanded now as we started to do a number of family studies, but I will show you the initial data that we had from a year ago looking at five patients with ulcerative colitis, Crohn's, and normals. The antigen we used was a novel antigen, keyhole limpet hemocyanin which people generally would not have been exposed to unless they were in our laboratory. We fed these people KLH daily for five days, rested them, then fed them for another five days, and then we drew blood sequentially through this period and immunized these people sequentially after the feeding then gave them a booster immunization. We measured both T cell proliferative responses and serum antibody responses. Here are the results. We used a very low-dose of KLH, because of studies from Wiener's group and actually from a number of groups have suggested that low-doses of antigens would be ones where you would induce these regulatory types, in this case Th3 cells that are activated by low-dose antigen administration.

So here is the normal individual's baseline responses, the T cell proliferative response after feeding - there is no response and post immunization one and two (I don't have two on the slide) - you basically see no response to the antigen. In contrast if you do the same study in patients with Crohn's disease after the feeding you actually see a boost in the T cell proliferative, it looks like these people have been immunized by oral feeding. It is not huge immunization to KLH, but after the oral immunization these people already make a booster type of response. It looks like oral feeding has actually primed a response to the KLH. The story is a little bit different in ulcerative colitis in that you do not see the immune response to KLH after feeding but after the first SQ immunization you see what looks like a booster response to the KLH. So these people obviously were primed by the feeding and generated this booster response. The antibody data although, clearly quite variable did, correlate with this tolerance/non-tolerance events that we saw in the different groups.

Okay, so what does this tell us? This tells us that there are defects in these regulatory mechanisms that occur in inflammatory bowel disease that we would like to equate with the defects that we see in the expression of this CD8 binding molecule gp180 and that if this is the case, the way in which patients with inflammatory bowel disease control responses to luminal contents are quite different than patients who are otherwise normal or healthy. Okay, so in the last five minutes, I would like to actually take you through some of the therapeutic implications of this and really start to underscore some of the similarities and differences in the therapeutic approaches between patients with Crohn's disease and patients with rheumatoid arthritis.

You all have seen this slide, this is a nauseating slide that talks about the balancing act between pro and antiinflammatory factors. This has been a pharmaceutical industry's dream because they can pick their molecule of choice and either try and block it such as occurred with TNF or IL-12 or alternatively add a regulatory cytokine that will then tip the balance back into a more steady state or homeostatic state that will allow for balance of both pro and antiinflammatory factors. Well the molecule that was targeted with the most success was to tumor necrosis factor. There was a lot of rationale for targeting tumor necrosis factor because it has effects on a wide variety of cells and most of these result in increased inflammation, tissue remodeling, compromising barrier function, etc, all the things that we see in patients with Crohn's disease. So what was generated initially by Centocor and by Celltech in England was a series of either chimeric antibodies which is what you see here initially called cA2 now called Infliximab or the trade name Remicade, which was a chimeric antibody which was about 25% mouse Ig or Celltech generating a more humanized antibody which is about 10-15% mouse and linked to IgG4 as opposed to IgG1, this antibody was infused into patients with Crohn's disease and actually rheumatoid arthritis beat Crohn's to the punch but the data was more impressive in the early Crohn's study. This is actually from the second study, the multicenter study that was published by Steph Targan and seven other centers where steroid-dependent Crohn's disease patients were given a placebo, 5, 10, or 20 mg. They were followed for a period of 12 weeks and asked whether they had a clinical response, which was a reduction in the disease activity index much like you guys have for rheumatoid arthritis or lupus, and then whether they actually went into remission. There are strict criteria for remission induction meaning that the activity index had to reach a level that was lower than 150.

You probably have all seen this data because every drug rep is throwing these in your face, but this tells us a couple of things. First of all there is a dramatic reduction in the activity index whether you use 5, 10, or 20 mg per kg in yellow, green, or pink and the results were quite dramatic within a very short period of time. So this is a rapid acting agent. We had not seen that in Crohn's disease. The other thing is that about a third of the patients went into remission, something that had not been described previously for patients especially in steroid dependent Crohn's disease. This was an impressive finding but the more impressive finding was looking at the longevity of this response whether we needed to give this drug once or, more than once. The second phase of this trial, which was published by Rutgeerts in Gastroenterology about a year and a half ago was taking those individuals who had initially responded to this Infliximab infusion and then randomized them to receive either placebo or Infliximab every eight weeks. Then monitoring them over a 44 to 48 week period.

Here are the results of that trial. This is the initial response - the patients by definition had to have responded. They all responded by eight weeks, a very nice response, that started to drift off by 12 weeks. At 12 weeks they were randomized to re-treatment with placebo or Infliximab. What you see here is two things. One is that the patients in this upper yellow or green line - I am color blind, I apologize - is the fact that you can maintain these patients in clinical response for a very protracted period of time by giving them q8week infusions. The other - to my mind the more impressive finding - is that the single infusion back here on day 0 resulted in, all the way out 44 weeks, a 40% clinical response rate in patients who were receiving placebo in the re-treatment phase. So this can induce very long lasting responses. The most recent trial was completed, and actually I was told the results were supposed to be available last week but they did not quite make it, was the ACCENT II trial that was done under the auspices of Centocor. This actually asked how often patients need to be re-infused, every four weeks, eight weeks, twelve weeks, etc. whether this was going to be a standard of care. Hopefully, these data will be available soon.

The other impressive thing which has came out of this trial, out of the Rutgers trial was the fact that there was an inflammatory measure C-reactive protein which dropped dramatically. TNF regulates CRP production, and that in these patients who were re-infused, down here on the bottom yellow line, the CRP level remains quite low, where as in those patients who were given placebo the CRP started to slowly rise back up to the pretreatment phase. Although, again I want to stress that this took a long period of time.

The other bit of evidence that we have that this was a drug unlike what we had previously used for Crohn's disease was the rapid induction in mucosal healing. Colonoscopies do not generally correlate with activity disease and Crohn's disease but this finding you see here, this marked inflammation here on the left, pretreatment, and this essentially normal restoration of mucosa four weeks after a single infusion was dramatic. It is sort of like what you see in RA as the reversal of periosteal resorption.

The other important area in which this hobby has been more successful is in treating patients with fistulizing disease. This is just a study, a multi-center trial that was done by Dan Present at our institution and the responses in fistulous disease were just as dramatic if not more dramatic in the patients treated with anti-TNF. Closure of all fistula, and these patients had between 1 and 10-15 fistula a horrible complication of Crohn's disease, had a dramatic response in 55% of the patients. A picture is worth a thousand words so here it is, pretreatment on the left, two weeks you are already see closure, ten weeks it is healing over, by eighteen weeks it is completely closed, very rapid response.

So, is this a good drug? It seems to be dramatic in a lot of patients with Crohn's disease and it is becoming more and more a mainstay of therapy in our patients. At Sinai we try and avoid the use of steroids. This has been a good way to try and do that but is it the be-all and end-all? The answer is we do not know. We also do not know, and that is why I threw in these two slides on safety, the long-term safety profile. It is amazing to me that over 60,000 patients have been infused from the Crohn's side with Infliximab since it was approved over two years ago. That is a scary number and it also tells you that probably the drug is not being used appropriately by gastroenterologists around the country, but in fact those numbers give us a little bit of comfort in terms of safety.

What have we seen so far? We have seen infusion reactions. I am sure you guys have seen these. These seem to be easy to treat through. We have seen some development of autoantibodies, which may be more of a factor in patients with RA than it is in Crohn's disease. There is only one patient who has developed a lupus-like syndrome who was treated for Crohn's disease. Generally this has not been a major issue. The issue that has been of concern and I am sure it is of concern for you guys, is the potential for lymphoproliferative diseases. In fact that was a report just two weeks ago about a NK lymphoma in a patient treated with Infliximab who had Crohn's disease, although, the details of how long he had the disease, how many infusion he had, has not been released by the company. There are a number (that actually came predominantly from the RA trials) of lymphoproliferative disorders that were described and there was only one case in the Crohn's study. Patients with Crohn's are more predisposed to developing lymphoma than the general population. I know that this is true for rheumatoid arthritis but certainly this is something we have to keep in the back of our minds as we are treating these patients with impunity and using these types of agents.

I will stop there and just remind you of how this paradigm and this therapy actually came to past with the use of agents which focused on this see-saw balance beam. The ability to target either the underproduction or over production of some of these factors to allow one to restore a balance and coexist with their external environment and ultimately to allow no disease to be present. I will stop there and I will entertain any questions. Thank you very much.

Question and Answer Session Following Grand Rounds:

Stephen Paget, MD: That was a superb presentation and obviously very similar to some of the presentations we have in rheumatoid arthritis as well especially with regard to the anti-TNF medications. Similar questions come up in different diseases. Obviously epithelial healing is different than joint erosion healing but clearly in both disorders these drugs work well. Two questions with regard to your molecule and the eventual development or lack of development of CD8 suppressor cells to be assured that it is not just a secondary phenomenon. Do you find that in these people who are treated who have an excellent response in Crohn's disease for example that the basolateral presence or absence is found or not found relative to support you hypothesis? Secondly, as we always discuss in rheumatoid patients why do 40 or 50 percent of patients not respond? Does that also have anything to do with your molecule?

Lloyd Mayer, MD: We have actually looked probably at about 200 patients with Crohn's disease from areas that are both involved and uninvolved and it is incredibly consistent not finding this molecule, at least in the intestine. The important thing that I want to stress is that this is from areas where there is no inflammation as well, so we do not think that this is a secondary phenomenon just relating to the inflammatory disease. The twin studies would also suggest that there is a defect in gp180 expression but this is not the cause of Crohn's disease, it may be another one of the contributing factors. If you treat patients - we did not think this was going to happen and it did not - with Infliximab and then looked pre and post treatment as to whether gp180 expression was up regulated, we still do not see it. So, we think that this is an inherent defect in the expression. We actually think in Crohn's disease there is a defect in glycosylation of CEA that does not allow this molecule (gp180)to be expressed and nothing we can do is going to bring that back unless we define which glycosylation enzyme if defective. We potentially might use that to restore expression and maybe that would help in controlling the inflammation. In terms of your second question, why 40% of the patients do not respond, that is the case I think, in any one of the colonic inflammatory diseases. You usually see that two-thirds respond and one-third does not respond and that it is never the same two-thirds or one-thirds that respond and does not respond. So clearly there are subpopulations of these diseases. IBD is not a single disease, hence the spectrum all the way from ulcerative colitis to Crohn's disease. There are multiple factors that probably come into play that gives one patient the substrate to allow them to respond to therapy X whereas that person may not respond to therapy Y and somebody who had just fistulous disease lets say or only ileal disease would respond. So, I think we need better markers in all of these chronic inflammatory diseases to really - Steph Targan has used the term reagent grade patients - really define these patients on a molecular basis and define the nature of the defects. Then we can pigeonhole them into what therapies might be best for that specific subset of disease.

Leonard Meiselas, MD: Lloyd many years ago, and you sort of referred to it when you put the word diet up there, there was a physician in Brooklyn who treated Crohn's disease with making sure those patients had no dairy food whatsoever and it seemed to have helped them clinically. What about diet? You did not mention diet being an issue and if it is what kind of an issue could it be?

Lloyd Mayer: I think the consensus at this point is that diet may be a contributing factor in terms of the symptoms but has no role in terms of inflammation, and actually I think those mouse models really emphasize that. If the mice are eating the same exact diet as their wild type litter mates and they are raised in a germ free environment they do not develop one ounce of diarrhea or inflammation. Yet they are bombarded by all of these potentially complex proteins and carbohydrates so dietary antigens do not play a role. In terms, of dairy products, there is a well-known association, especially with Crohn's disease and secondary lactose intolerance. I have lactose intolerance and I do not have IBD. I know that if I drink a glass of milk I am going to have abdominal cramping and eventually I am going to wind up in the bathroom. Because lactose is a very sensitive enzyme, it gets knocked off epithelium very easily and if this happens and if you are going to drink a glass of milk you are going to have the problems that any genetically lactase deficient individual would have and I think that is probably it. You are not going to give a patient with obstructive Crohn's high fiber diets because that is going to cause more problems. So, I do not think that diet is really playing a major role and I think that other than the British who still feel that Rice Krispies or Corn Flakes or refined sugars cause Crohn's disease, you will find anybody else that really buys into that anymore.

Stephen Paget: Recent studies have demonstrated the effectiveness of growth hormone in inflammatory bowel disease. Where does that fit into this paradigm that you have discussed?

Lloyd Mayer: That is actually an interesting study and we are involved in submitting a multi-center grant on that now. Alf Slonim is a pediatric endocrinologist from North Shore, who made the fortuitous observation that in a couple of kids he had with growth retardation from Crohn's disease, when treating them with growth hormone, their GI symptoms got better. From that he did a study that they published in The New England Journal in June which suggested that there was a marked reduction in symptoms. There are a lot of problems with that study. It is actually surprising in some sense that it got published in the New England Journal. I guess the novelty was enough to get it in there. But growth hormone has some immunomodulatory affects. The literature is sort of weak and we have been doing this now in preparation for submitting this grant but there are a number of studies that have looked at altering the balance of Th1 and Th2 by adding growth hormone in vitro. This may be one mechanism where by it works. Growth hormone also has effects on somatostatin, which is an incredibly negative factor in terms of controlling inflammation. Somatostatin knocks off T cell proliferation, antibody production, so people think that if you put too much growth hormone in, that secondarily you are going to get increased somatostatin that is going to shut down the inflammation. You know what? The mechanism is totally up in the air and that is actually why we are submitting this grant.

Stephen Paget: They had implied the issue of mucosal integrity with the lack of proteins being able to go across.

Lloyd Mayer: Dusty Ziegler at Emory has suggested that because he has done studies in short gut syndrome where growth hormone actually seems to improve the integrity. The problem was that in the studies that Alf did with patients that there was no evidence that there was any improvement in permeability. I think that would be nice if that were the mechanism. It would be, you know, easy to ascribe it to at least that but that has not turned out to be the case.

^ Back to Top
Request an Appointment